Yu-Tsung Hu

Title: Innovations of Yu-Tsung Hu in Video Stabilization Technology

Introduction

Yu-Tsung Hu is a prominent inventor based in Changhua County, Taiwan. He has made significant contributions to the field of image processing and video stabilization. With a total of 10 patents to his name, Hu's work is recognized for its innovative approaches and practical applications.

Latest Patents

One of Yu-Tsung Hu's latest patents focuses on a training method for video stabilization and an image processing device utilizing this method. The proposed technique involves several steps, starting with the reception of an input video containing low dynamic range (LDR) images. These LDR images are then converted into high dynamic range (HDR) images using a first neural network. Following this, a feature extraction process is conducted to obtain features based on both the LDR and HDR images. A second neural network is trained for video stabilization, generating stabilized HDR images in a time-dependent manner. This process minimizes a loss value based on a loss function that depends on the extracted features. Additionally, an HDR classifier is constructed to classify the stabilized HDR images, generating a reward value that further influences the loss value.

Career Highlights

Yu-Tsung Hu is currently employed at Novatek Microelectronics Corporation, where he continues to develop innovative technologies in the field of image processing. His work has garnered attention for its potential applications in various industries, particularly in enhancing video quality and stabilization techniques.

Collaborations

Some of Yu-Tsung Hu's notable coworkers include Tsui-Chin Chen and Yu-Shu Liu, who contribute to the collaborative efforts in advancing their projects and research.

Conclusion

Yu-Tsung Hu's contributions to video stabilization technology exemplify the innovative spirit of modern inventors. His work not only enhances the quality of video processing but also paves the way for future advancements in the field.